FeaturesPricingUse CasesIntegrationsCompare
Original research, October 2026

Do B2B websites change for paid-search visitors? A study of 300 startup homepages

Almost never. Of the 264 B2B startup homepages we could assess on October 25, 2026, 1 (0.4%) served a different headline or page title when the URL carried Google paid-search UTM parameters, and none repeated the search term. When we rechecked that one site, its headline also changed between two requests without UTM parameters, so we found no homepage that demonstrably adapts its server HTML to the paid-search visitor. Of the 44 that load an ad-platform pixel, 44 (100%) showed the paid-search visitor exactly the same headline and title as everyone else.

Published
October 25, 2026
Author
Andreas Hatlem, founder of GetIntent (Getia AS)
Sample
300 randomly drawn Y Combinator B2B companies, 281 fetched
Data collected
October 25, 2026, single run

Key findings

  1. 1

    1 of 264 assessable homepages (0.4%) changed their first headline or page title when fetched with utm_source=google, utm_medium=cpc and a test utm_term. 0 repeated the test search term anywhere in the visible text. On a recheck the same day, that site's headline also changed between two plain requests without UTM parameters, so the difference is more likely headline rotation or an A/B test than a response to the campaign.

  2. 2

    45 of 281 fetched sites (16%) load an ad-platform pixel (Google Ads, LinkedIn, Meta or Microsoft Ads), a sign that they measure paid traffic. 44 of the 44 assessable ones (100%) served the paid-search visitor the same headline and title as everyone else.

  3. 3

    71 sites (25.3%) have a form in the homepage HTML. 1 of them (1.4%) capture UTM parameters or the Google click ID in hidden form fields, so campaign data rarely reaches the CRM with the lead.

  4. 4

    26 sites (9.3%) mention UTM parameters in inline scripts, typically to store them for analytics or attribution rather than to change the page.

  5. 5

    7 sites (2.5%) load a known personalization or experimentation script such as Mutiny, Intellimize or VWO. These tools can change copy in the browser after the page loads, which this study does not observe.

  6. 6

    21.2% of assessable homepages use at least one generic call-to-action label such as “Get started” or “Learn more” in the page body, and on 42% our parser found no button-style call to action in the body HTML. 25.6% have no canonical tag and 4.3% no meta description.

What the homepages do with campaign traffic

Share of sampled homepages per measure. Each bar states its own base, because some measures only apply to pages we could assess or to pages with a form.

  • Load an ad-platform pixel16% (45 of 281)
  • Load Google Tag Manager29.9% (84 of 281)
  • Use a generic CTA label21.2% (56 of 264)
  • Mention UTM parameters in inline scripts9.3% (26 of 281)
  • Load a personalization or experimentation tool2.5% (7 of 281)
  • Forms that capture UTM or click IDs1.4% (1 of 71)
  • Change headline or title for a paid-search visitor0.4% (1 of 264)
Bases: tools and UTM scripts use all fetched sites; headline, title and CTA measures use assessable sites; hidden UTM fields use sites with a form.

Marketing tools found in the homepage HTML

Detected by matching known script URLs and tag IDs in the HTML of the 281 fetched homepages. Tools loaded only through a tag manager are not detected, so these shares are lower bounds.

ToolSitesShare of fetched sites
Google Tag Manager8429.9%
Google Analytics 46623.5%
PostHog3813.5%
Google Ads tag248.5%
HubSpot207.1%
LinkedIn Insight Tag186.4%
Meta Pixel176%
Segment62.1%
VWO41.4%
Hotjar31.1%
Microsoft Ads UET31.1%
Intellimize20.7%
6sense10.4%
Clearbit10.4%
Marketo10.4%
Mutiny10.4%

What this means if you pay for clicks

A visitor who clicked an ad for a specific search term almost always lands on the same page, with the same headline, as every other visitor. Matching the page to the query (message match) is mostly left undone, even by companies that track their ads closely.

There are several ways to close that gap: a dedicated landing page per ad group, dynamic text replacement, or AI personalization such as GetIntent, which rewrites the headline and copy of the existing page for each campaign and search term. Whichever you choose, measure it against the original page with a control group before you roll it out.

Methodology

Population: every company in the public Y Combinator company dataset published by yc-oss (yc-oss.github.io/api/companies/all.json) with industry “B2B”, status “Active” and a website, deduplicated by hostname. That is 2265 of the 6336 companies in the dataset on October 25, 2026.

Sample: 300 companies, drawn at random with a seeded shuffle (Mulberry32, seed 20261025) after sorting by YC company slug, so the same dataset snapshot always gives the same sample. Every sampled company and website is listed in the CSV below.

Procedure: each homepage was fetched twice with the same rules as our free personalization audit: once as listed, and once with utm_source=google, utm_medium=cpc, utm_campaign=getintent-personalization-audit and utm_term=getintent-audit-probe added. Requests identified themselves as GetIntentAudit, followed at most 5 redirects, read at most 1.5 MB and stopped after 25 seconds per site, with 6 sites fetched in parallel. The HTML was parsed without running JavaScript. The run took place from Oct 25, 2026, 3:02 AM to Oct 25, 2026, 3:02 AM UTC.

Definitions

Changes for a paid-search visitor
The first H1 or the page title differs between the two fetches, or the test search term appears in the visible text of the second fetch.
Assessable homepage
Fetched successfully and not rendered only in the browser. 17 fetched homepages had no H1, under 200 characters of visible text and at least one script, and are left out of the headline, title and CTA measures.
Ad-platform pixel
Google Ads tag, LinkedIn Insight Tag, Meta Pixel or Microsoft Ads UET found in the HTML.
Captures UTM in forms
A form contains a field whose name or ID contains utm_, utm-, utmsource, utmcampaign or gclid.
Generic CTA label
A button or button-styled link outside the navigation, header and footer labelled with a generic phrase such as Learn more, Read more, Get started, Sign up, Submit, Contact us or Start.
Personalization or experimentation tool
A script from Mutiny, Intellimize, Dynamic Yield, Optimizely, VWO, AB Tasty, Convert or GetIntent found in the HTML.

Limitations

  • We read server HTML only. Copy changed by JavaScript after the page loads is not observed, so sites that personalize in the browser count as unchanged.
  • We tested homepages. Many companies send paid traffic to dedicated landing pages, which may match the ad even when the homepage does not.
  • One pair of requests per site, from one location (Norway), on one day. Pages that vary by country, test group or time of day may look different elsewhere. A site that rotates its headline on every request can therefore count as changing for the paid-search visitor; we rechecked the one site that changed, and it does rotate.
  • Tool detection only finds scripts referenced directly in the HTML, so tool shares are lower bounds.
  • Young Y Combinator B2B companies are not all B2B companies. The results describe this population, not the whole market.
  • 19 sampled sites could not be fetched (unreachable, address blocked by our network safety rules, or an HTTP error) and are excluded from every measure.

Check or reproduce the results

Every sampled website is in the CSV with its measured values. To check any row, run that URL through the free personalization audit: it uses the same analyzer and the same test parameters. To repeat the whole study, apply the population filter, sorting and seeded shuffle described above to the yc-oss dataset; our study script is available on request at [email protected].

How to cite this study

GetIntent (Getia AS), “Do B2B websites change for paid-search visitors? A study of 300 startup homepages”, October 25, 2026, https://getintent.co/research/b2b-utm-response-2026

Last updated October 25, 2026. Data collected October 25, 2026.